Consistent Antarctic sea ice concentration biases from passive microwave satellites resolved via ICESat-2 altimetry

Observational products that convert passive microwave (PM) emissions from the Earth's surface into sea ice concentration (SIC) are critical and long-standing tools for understanding polar change, both for understanding changes to overall sea ice area (SIA) and evaluating and improving global climate models. However, the algorithms and sensors used to convert the PM signature into an estimate of SIC can lead to differing estimations of Earth's sea ice cover and its change over time. Here, we show that in Antarctica, where sea ice loss since 2016 has exceeded 2 million km2, inter-algorithmic and inter-sensor biases lead to differences in overall SIA of up to 1.7 million km2, and of Antarctic marginal ice zone area by up to 4.7 million km2. Using 6 years of coincident data from the ICESat-2, we show this bias comes from an underestimation of SIC in thin ice zones by commonly-used SSMI/S-based products. These biases are readily parameterizable as a function of SSMI/S sea ice concentration, leading to a unified SIC product that resolves persistent biases in PM-SIC data.